1
Influence month of birth on success
football players
Thomas Res 10454918
Supervisor: Ron van Maurik
15-07-2016
Bachelor Thesis
University of Amsterdam
2
Statement of Originality
This document is written by student Thomas Res who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3
Contents
1. Introduction
4
2. Literature review
5
3. Data
8
4. Methodology
12
5. Conclusion
15
References
17
Appendix A
18
Appendix B
19
Appendix C
20
4 1. Introduction
Imagine identical twins where the oldest is born just before new year on the 31th of December and the youngest is born just after new year on the first of January. Both have the same genes, same education and are born with the same sport talent. Yet, the youngest has far better chances of becoming a professional sports athlete. This is due to the relative age effect. A lot of research has been published about the relative age effect in sports over the last decades. According to the existing literature about the relative age effect, football clubs should be aware of this phenomenon by now and change the way of scouting. No longer scout the best players, but the players with the highest potential. Still, in selections of sport teams players born in the early months of the selection year are overrepresented.
Football is a big business. Clubs want to make money, so they are able to buy players of high quality. If a club can increase their budget, they can afford signing high quality players and are more likely to gain success. It is in the best interest of a football club to have players in the squad who represents high value. Most football clubs have more players born in the early months of a selection year than players born in the last months (see table 1), while there is no evidence that players born in those early months have more potential than players born in the last months. This relative age effect has been known for several decades now. But apparently players born in the first quarter are still overrepresented. Is this a sign of bad investment or are players born in the first quarter worth more than players born in other quarters? What is the impact of month of birth on the success (quality) of
football players in the Dutch League?
Table 1. Month of birth in youth academies professional Dutch football clubs
Bron: Tussendelinies.nl
To measure the rate of success of an individual football player his market value will be used as an indicator. If a player has skills and can add value to a team in order to win prices, teams have to pay a price in order to contract him. The more qualities a player has, the more money his club can ask. The market value of a player willbe a good indicator of the qualities of a player.
To investigate the impact of month of birth on the market value of football players in the Dutch League data of all football players who were on the pitch during any League match in the Dutch first division in the season of 2015/2016 will be gathered. Then a regression analyse will be performed. Variables that could influence the market value of a player will be included. The age of a player, his position on the field and his month of birth. A regression analyses will be done to show the influence of all these variables and to see if the month of birth of a player does have an influence on their market value.
5 In paragraph 2 the existing literature will be discussed. In paragraph 3 the dataset will be explained and in paragraph 4 the results will be shown. In paragraph 5 the results will be discussed and a conclusion will be drawn.
2. Literature review
To become a successful sport athlete, several factors play a role. Including gender, training, nutrition, family background and social cultural influences (Nakata & Sakamoto, 2011). In the last two decades, the month of birth is also considered a relevant factor.
In the 1980s this effect in sport was first noted in Canadian ice hockey. Grondin, Deshaies & Nault (1984) and Barnsley, Thompson & Barnsley (1985) concluded that success as a hockey player is related to the month of birth. They suggested that this was due to the fact that if children are categorized based on their age, which is the case in sport, children born immediately after the selection date are approximately 12 months older than children born just before this selection date. The older children in the age group have a developmental advantage over the younger children in the same age group. The children born in the first months after the selection date are in general bigger, stronger and better coordinated than their teammates who are born just before the selection date. As a result, when these children play sport together, in general the older children perform better. Because they are doing better, the older children achieve more success and receive greater awards. The younger children are more likely to experience frustration and failure. This could lead to a lower self-confidence and self-esteem. As a result of these experiences, according to Barnsley and Thompson (1988), they may quit doing this sport and so their potential talent will not be exploited. After the result from Barnsley, Thompson and Barnsley in 1985, there has been done many research on the relative age effect in several sports. In football, worldwide sport number one, the relative age effect has been researched all over the world. This has been done by looking at the population of month of births of football players in all sorts of competitions. To investigate whether the population is equally distributed most researchers used a Kolmogorov-Smirnov test or a chi-square test. All players of the sample will be divided on the basis of their month of birth in quarters. Players born in the months January till March will be placed in quarter one, players born in the months April till June will be placed in quarter two, etcetera. Then a one sided Kolmogorov-Smirnov test is used to test if the distribution of the sample fits the chosen distribution, in this case a uniform distribution. Or a chi-square test is performed. With the chi-square test an expected value will be calculated for all possibilities. On the subject of the relative age effect, the months of birth will be examined, so for every month the expected total amount of players born in that month will be calculated. The chi square test will be based on a uniform distribution. For example January has 31 days, so the chance of a player being born in January = 31/365.25 = 0.085. The expected value for January is then 0.085 * total amount of players. The expected values will be compared with the observed values and if there is a significant difference, the conclusion will be drawn that the month of birth of the players is not uniform distributed. Skewed birth-date distributions have been revealed in favour of the individuals born in the first months of the selection year all over the world. Verhulst (1992) showed the relative age effect in the first and second professional divisions of Belgium, the Netherlands and France. Dudink (1994) reported a significant relative age effect in all top four leagues in England. Musch & Hay (1999) found a strong relative age effects in Germany, Japan, Brazil and Australia. Bäumler (1996) showed the effect in the highest professional division in Germany. However, Bäumler found
6 that the relative age effect among the youngest players (18-20 years) was much stronger than the relative age effect among the oldest players (33-35 years). In fact, the relative age effect on the oldest players was not significant at all. According to Bäumler this was evidence that the physical advantage of the players born early in the selection year decreases over time.
Most researchers argue that the cut-off dates are the only factor underlying skewed birthdate distributions. Musch & Grondin (2001) stated that this claim must be defended against all other possible alternative explanations. Wendt (1978) showed that seasonal circumstances could influence the child’s development. For example, children go to different phases of motor learning. Warm weather during an important phase, could have a positive influence on the development of a child’s sport-related skills. According to Wendt children born in certain months of the year can profit from the fact that their critical sensitive phases are during the summer rather than during the winter. Therefore it could be that these seasonal circumstances lead to a skewed birthdate distribution. To investigate this potential factor of the skewedness of the month of birth distribution in football, Musch and Hay (1999) compared the relative age effect in German football with the relative age effect in Brazilian football. Both countries have a highly developed soccer system and used a cut-off date of August 1. However, climatologically Germany and Brazil are exact opposites. If there would be a 6-month shift in the pattern of birth dates, this would suggest that there is an influence of season or climate. It turned out the pattern of birth dates were the same in both countries, so this suggests that the cut-off date is the reason of the skewed birthday distribution.
Wendt (1974) and Barnsley, Thompson & Legault (1992) pointed out that the findings on the relative age effect in sports had to be separated from the possible influences of the cut-off date for school entrance. Almost every public school uses cut-off dates to decide in which class a child will be placed. Doornbos (1971), May & Welch (1986) and Grondin, Proulx & Zhou (1993) showed that the relative younger children in class have more academic problems than their older classmates. In countries were the cut-off dates of school entrances are the same as the cut-off date which is used in football, the advantages of the older children at school may have an confounding effect in sport according to Musch and Hay (1999).
Musch and Hay (1999) investigated this by looking at the distribution of birthdates in Australian soccer. The Australian football bond changed the cut-off date from January 1 to August 1 in 1988, while the cut-off date for school entrance did not change. Musch and Hay found that there was a corresponding shift in the player’s birthday distribution 10 years after the change, so this provides strong evidence for the existence of a relative age effect, solely caused by the cut-off date.
Condition
Musch and Grondin (2001) state that competition is a necessary condition for the relative age effect to exist. If there are as many children willing to play a certain sport in a certain category to fill up one team, there is no reason for relative age effect to exist. Every player will be placed in the best (and only team), no matter the skills. If the demand to play a sport is very high, a selection has to be made who will play in the first squad. The relative age effect is then more likely to occur. This condition was found by Grondin, Deshaies & Nault (1984). Their results showed that the relative age effect in ice hockey was stronger in cities with more ice hockey players. They also tested the distribution of month of births of volleyball players in Canada. Volleyball is less popular in Canada. For volleyball
7 teams in school, they did not found a relative age effect at all. Because football is a popular sport worldwide, the relative age effect isstrong.
Physical development
Malina (1994) showed that physical development is a factor of success in sport. Malina revealed that a one year difference in age can make a big difference in weight and stature, for young children. Grondin and Trudeau (1991) investigated the relative age effect in the NHL, ice hockey. They found that 55% of the forwards were born in the first 6 months of the year. But more than two third of the goalkeepers were born in the first half year. Grondin and Trudeau suggested that these findings could be explained by the fact that goalkeepers wear the heaviest equipment. Physical strength is of great importance for a goalkeeper in ice hockey. Children mature in different time, some children have early maturation and some children experience late maturation. Malina (1994) found that in every sport, except gymnastic, early maturation is an advantage. According to Malina, gymnastic is an exception, because of the extremely selective criteria in this sport, including selection at an early age for small body size and physique characters associated with later maturation. The relative age difference is largest at a young age. So the relative age effect should decrease together with an increase in age according to the factor of physical development.
Psychological variables
Psychological variables should also be considered as a factor of the relative age effect (Musch and Grondin, 2001). Children differ not only physical, but also in their psychological maturity. According to Feltz & Petlichkoff (1983), Roberts, Kleiber, & Duda (1981) and Vallerand, Deci, & Ryan (1988) children with a high perceived competence show a higher intrinsic motivation and enjoy their sports more than children with a low perceived competence. Children will become more aware of their own competence as they become older. So the drop out of less competent children will increase as these children become more aware of their competence. This means that the relative age effect by
children would increase together with an increase in age according to this psychological factor. After a certain point the psychological effect will not lead to an increase in the relative age effect anymore. The perception of competence of the players will not differ much from their actual competence once they become of age.
Experience
Another factor that contributes to the relative age effect is the experience (Musch and Grondin, 2001). A player who is a few months older is not only physical stronger or psychologically, but simply has more life experience and has had more hours of training than his younger equivalent. So purely based on his experience his chances of being selected for the first team are higher. As a result, playing in the first team is associated with receiving better coaching and playing against stronger opponents which stimulates the quality of the players. Playing in the first team will likely lead to an increase in motivation, because it gives more prestige and matches will be played for a bigger
audience. The difference in experience is relatively largest at young age. The factor of experience will become of less importance as age increases. The relative age effect should therefore decrease if age increases, looking at the contribution of experience.
8 3. Data
The dataset contains all the players who played at least (a part of) one League match of the
2015/2016 season in the Eredivisie. Player who only played during post-season play off games and/or Cup matches are not included. The data is constraint from the website whoscored.com and double checked with the website transfermarkt.com. Players who left during the season were only on the whoscored website. For every player their market value is retrieved from transfermarkt.com. This is an international website which is considered an accurate way to discover how much a player is worth. The website determines this value using the following criteria:
Qualities and potential
The quality and potential of a player is the biggest factor to determine his market value. The higher a players football skills are, the higher the price other teams will have to offer to buy this player. If a player shows potential and the market thinks he is ready for the next step to a bigger club or a better competition, his transfer value will increase.
Performance
The statistics of a player are a factor. For every position and roll different statistics are important. The simplest statistic for a striker for example are his total goals scored. The amount of injuries a player affects his market value, a player who plays every match is worth more than a player who is a substitute and so there are many statistics which will influence the market value.
Level of club and competition
The international status of a club and competition plays a roll. A player can perform outstanding, but if his performances are in a low competition for a small club, his market value will be lower than another player whose performance is at the same level, but at a bigger club and/or competition.
Age
If a player with an age of 21 and a player who is 34 years old have the same qualities and all other variables are equal, the youngest player will be worth more. The club can profit longer from the services of this player and the residual value for a young player is much higher than the residual value of an older player.
Position
The position of a player has its influence on the market value. A defender is in general worth less than a striker. If a player can play at different positions, it increases his market value as well. Variables
The qualities, potential and the performance of players can’t be measured objectively. These criteria are too complex and can’t be included in the regression analyses. Only the players from the Dutch highest league are included in the data, so all players play in the same competition. Therefore the competition won’t have an effect on the different market values. For every player their club, age and position is added to determine the impact of these factors on their transfer value. The date of birth of players is also added to determine if the month of birth has an impact on the market value.
9 The dataset contains 458 players who had at least one appearance in the Dutch league in 2015/2016. 150 players are born in quarter 1, which is equivalent to 32,8% and 78 players (17,0%) are born in quarter 4 (see table 2). Of the 458 players, 284 are Dutch (see Appendix).
Table 2. Total players by quarter and average age Quarters Total players Average age
Q1 150 23.77
Q2 122 23.89
Q3 108 23.88
Q4 78 25.13
For the market value the ln of the market value will be used. The market values vary from 50.000 to 15.000.000, so the natural log will be used to decrease the difference and give a better picture (see figure 1).
Figure 1. Distribution of natural log of the market value
The club of the players is available on transfermarkt.com. Players who played matches for two different teams in this season initially show up twice in the data. Their market value is determined based on their current team, so the player in combination with his former club will be removed. Only the player in combination with his current club will remain part of the data.
For all players the market value on the 18th of May 2016 will be used. The competition was finished
on the 8th of May and a few updates on the transfer values were done by transfermarkt.com the days
after this last match day. For every player their age on the 18th of May will therefore be used. This
0 .2 .4 .6 D en s it y 11 12 13 14 15 16 lnMV
10 variable is obtained by subtracting the date of birth of a player from the current date (May 18th).
Most players in the Dutch highest football league ended the 2015/2016 season with an age between 19 and 25, with a peak at 21 (see figure 2). The oldest player has an age of 35, the youngest player in the data has an age of 16. In order to make the difference in age more visible and easier to compare, for every player their age will be subtracted with 16. The adapted age of the youngest player will now be 0.
Figure 2. Distribution of age
The position of the players is obtained from transfermarkt.com and whoscored.com (see figure 3). Sometimes it can be hard to determine a players position. Some players are able to play at multiple positions and play different positions in different matches or even in the same match. In the cases the websites uses different positions for the same player, the position of transfermarkt.com is used. Whoscored.com determines the position of players once and during their whole carrier they will be considered to play in that position, unless their position has obviously changed. Transfermarkt.com evaluates a players position constantly and is therefore considered the more accurate one.
0 .05 .1 .15 D en s it y 15 20 25 30 35 Age
11 Figure 3. Amount of players per position
Players who have multiple nationalities (including the Dutch nationality) were assigned the nationality of the country for which they play international football. Once they have played for a country, they can’t play for any other country anymore according to the rules of the Fédération Internationale de Football Association (FIFA). They will be threated as a member of the country they are playing for. If these players never played an international match, they have been assignedthe Dutch nationality. The vast majority of these players, if not all of them, are born in the Netherlands, went to school in the Netherlands and enjoyed their football education at a football academy in the Netherlands. For the purposes of this essay, they should be considered Dutch.
Table 3 shows the clubs ranked on the total market value of all their players who played a part in at least one League match in the season 2015/2016 compared with their position in the League in 2015/2016. The total players used shows the amount of players who took part in the competition in season 2015/2016 by club. It adds up to a total of 473, which is more than the total players in de dataset (458). This is due to the fact that 15 players played for multiple teams. During the transfer periods in August or January they have been transferred to another club in the Eredivisie (see Appendix for the list of players who played for multiple teams).
Table 3. Total market value by club and total players used.
Market value position
Club Total
Players used
Total value League position Difference 1 PSV Eindhoven 23 116,55 mln. € 1 0 2 Ajax Amsterdam 25 116,50 mln. € 2 0 3 Feyenoord Rotterdam 24 61,55 mln. € 3 0 4 AZ Alkmaar 28 39,78 mln. € 4 0 5 Vitesse Arnhem 26 31,98 mln. € 9 -4 6 FC Twente Enschede 28 30,85 mln. € 13 -7 7 FC Utrecht 27 29,55 mln. € 5 +2 8 SC Heerenveen 22 26,48 mln. € 12 -4 30 153 129 146 Keeper Defender Midfielder Striker
12
9 FC Groningen 24 24,53 mln. € 7 +2
10 NEC Nijmegen 27 19,10 mln. € 10 0
11 PEC Zwolle 29 17,10 mln. € 8 +3
12 ADO Den Haag 24 16,65 mln. € 11 +1
13 Willem II Tilburg 29 16,55 mln. € 16 -3 14 Heracles Almelo 22 15,60 mln. € 6 +8 15 SC Cambuur-Leeuwarden 32 11,65 mln. € 18 -3 16 Roda JC Kerkrade 33 11,53 mln. € 14 +2 17 Excelsior Rotterdam 25 10,05 mln. € 15 +2 18 De Graafschap Doetinchem 25 8,88 mln. € 17 +1 Bron: transfermarkt.com 4. Methodology
A chi square test will be performed to indicate the existence of the relative age effect in the Dutch highest football League. The expected numbers of players born in every month will be compared with the observed numbers of players born in every month. As other researchers on the relative age effect have done, an assumption will be made that the possibility of a child to be born on a date is as high as any other date. The chi square test will be used to investigate whether the distribution of months of births in the Dutch League is uniform or not. If it turns out the distribution is not uniform, but has a skewedness in favour of the individuals born in the first months of the selection year the existence of the relative age effect will be confirmed.
With the chi-square test an expected value will be calculated for all possibilities. The months of birth will be examined, so for every month the expected total amount of players born in that month will be calculated. The degrees of freedom are equal to 11, the total amount of months minus one. The chi square test will be based on a uniform distribution. For example January has 31 days, so the chance of a player being born in January = 31/365.25 = 0.085. The expected value for January is then 0.085 * total amount of players. The expected value will be compared with the observed value and if there is a significant difference, the conclusion will be drawn that the month of birth of the players is not uniform distributed.
The difference between the expected values and the observed values leads to a chi-square value of 26.785, which correspondence to an asymptotic significance of 0.005, which is less than the
significance level α of 0.05 (see Appendix). The chi-square test shows that the months of birth of all the football players in the Eredivisie is not uniform distributed. The cut-off date which decides in what age category children will be playing football is in the Netherlands, and in almost every other country, the 1th of January. The distribution has a deviation towards the first months of the selection year, which indicates that the relative age effect is still present in the Dutch highest football League. Now a regression analysis will be performed to investigate whether the date of birth has an influence on the market value of a player. The dependant variable is the natural logarithm (ln) of the market value. The ln of the market value is used in order to decrease the big differences as the market value varies from €50.000 to €15.000.000. As independent variables the position, quarter of birth and (adapted) age will be used.
13 The club of a player also influence the market value. Playing at a top team will increase the market value of a player. There is more pressure to handle at a top team, more competition in the squad and the status of the club has a positive influence on the market value. Besides, a big team can simply ask more money. A club with a smaller budget will more likely accept a lower offer, because for them the offer is relatively much more.
But in the regression model the club of players can’t be included is a variable. The best players play for the best clubs, because they offer the most money and have the best chances of winning trophies. So in general players who play for the best teams have the highest market values. But that is not because playing for those clubs leads to a higher market value, it is because they are top quality players. And because they are top quality players, they are playing for the best teams.
14 Regression 1 shows only significant values for Age1 and Defender. An increase in age has a positive effect on the market value (see table 4), while the theory states that an increase in age should have a negative effect. A player will be less attractive because he has less potential of growth, his residual value will be less, he can perform fewer years for the club and in general his qualities will slowly decrease in time after he has reached his peak. The other variable that has a P-value less than the significance level α (0.05) is Defender. It has a coefficient of -0.35. This means that if a player is a defender, it has a negative effect on his market value. Defenders are in general worth less than players who play on other positions.
A second regression is performed to investigate the influence of age on the relative age effect. Some researchers have showed that the relative age effect is strongest at young age and that for the older players the relative age effect is not significant anymore. To see whether the age has an effect on the relative age effect the quarters will be combined with the variable Age1, which is the Age of the players minus 16.
The variable age on itself is not statistically significant anymore. There is no proof that the age of football players has a positive or a negative effect on their market value. The variable 1.Quarter1 shows the effect on the natural log of the market value for a player who is born in the first quarter of the year. It gives a value of -1.21 and is statistically significant. It means that if a player is born in quarter 1 it will have a negative effect on the natural log of the market value of 1.21. The variable Quarter1#c.Age1 is also significant and has a value of 0.14. This shows that if a player is born in quarter 1 his market value will increase if his age increases.
In figure 4 this is shown by a graph of the derivative of the natural log of the market value with respect to quarter1. The derivative of the natural log of the market value with age on the x-axis is now a linear function increasing with an increase in age. This shows that the relative age effect of players born in quarter one is decreasing as they become older. This function is increasing so much that players with an age of thirty or more are worth more than other players. This isn’t realistic. Therefore instead of the variable Age1 the log of Age1 will be used in regression 3. Now the
derivative with respect to quarter1 will be increasing, but because of the logarithm the increase will tone down. Figure 4 -1 -. 5 0 .5 1 1 .5 d er lnM V 15 20 25 30 35 Age derlnMV derlnMV
15 For players born in the second quarter there is a similar effect. The effect is smaller than for players born in the first quarter and only the direct negative effect of being born in the second quarter has a significant value. The negative effect of being a defender has barely changed and is still significant. In regression 3 the variable Age1 is replaced with the log of Age1. The derivative of the natural log of the market value with respect to quarter1 with age on the x-axis is not linear anymore (see figure 5). The relative age effect is still decreasing with an increase in age, but slowly converges.
Figure 5
5. Conclusion
The existing of the relative age effect in football has been proven by many researchers over the last decades. Barnsley, Thompson and Barnsley were the first who discovered this phenomenon in 1985 while they were studying the Canadian hockey leagues. They showed that there were significantly more hockey players born in the first quarter of the selection year than players born in the last quarter of the selection year. After this result, many research on the relative age effect was done and several researchers showed that the relative age effect exist in professional football all around the world. Some researchers showed that the relative age effect diminishes as players become older. The existence of the relative age effect is now known approximately twenty years. Professional
organizations of football clubs should be aware of the advantage of players born early in the year, but the relative age effect isn’t decreasing. Apparently clubs are still choosing the best players at that moment to play for their youth academy instead of the players with the highest potential.
An import aspect of a football club is the transfer value of their players. If they can realize an increase in their budget, new players can be signed to improve the squad. The market value of all players in the Dutch highest league will be examined to look whether the players born in the first quarter of the year have a lower transfer value than other players. Current literature shows that there are more football players born in the first quarter of the year, while there is no evidence why players born in the first quarter of the year are better football players than others. This could be a sign of
mismanagement, but it doesn’t provide evidence. If the transfer value of players born in the first quarter are significantly lower than the transfer values of other players, then there is evidence that
-2 -1 0 1 2 d er lnM V 15 20 25 30 35 Age derlnMV derlnMV
16 clubs are hiring too much players born in the first quarter of the year and should have a more equal distribution.
The dataset contains 458 players who had at least one appearance in the Eredivisie in the season 2015/2016. Of these 458 players, 150 were born in the first quarter of the year, 122 were born in the second quarter, 108 in the third quarter and 78 in the fourth quarter. A chi square test was
performed and showed that the distribution of the month of birth of the players is not equally distributed. The distribution has a deviation towards players born in the first months of the year, which proves the existence of the relative age effect in the Dutch league in the season 2015/2016. A regression analyses is performed with the natural log of the market value is the dependent variable. Regression 1 indicates that age has a positive effect on the market value. This can be explained by the fact that in the Dutch League many young players get a chance to play for the first squad. There are many youngsters who played only a few minutes this season and have a very low market value. A closer look on the data shows that there are 35 players who have a market value of €150.000 or less. Of those players 34 are 22 years old or younger. The reasoning behind this is if players become older than 22, they are no longer considered as talents whose time is still to come. If they have showed improvement and are now good enough, they play. If they are still not playing at their club, they want to go somewhere else. Besides, for the clubs these players become too expensive to be a reserve. For both parties it is best to look for another club and they go separate ways.
Regression 2 and 3 show that players born in the first quarter of the year have a lower market value than players born in order quarters of the year. This effect is decreasing as players become older. When players born in the first quarter of the year become 18 and are expected to be ready to play for the first squad, their advantage over their younger teammates is over. During their time in the youth academy they always could rely on their (physical) advantage of being a few months older. Now it becomes clear that some of these players are not as good as it seemed to be all that time. Subsequently these players will be rewarded with a low market value. They show less quality, so other clubs should be able to buy these players for a low price.
On the other hand, the players on the academy who are not born in the early months of the year, had to overcome an extra obstacle. They didn’t had the advantage if being older than all the other players, but still were invited to the academy. This means they compensated the lack of advantage of the relative age effect. These players are in general more talented than part of the players born in the first quarter. When they become of age and start playing for the first team, their talent and qualities remain visible. This leads to a higher market value.
This effect will decrease with an increase in age. Of some players born in the first quarter who made it due to their relative age advantage, it becomes clear they are not good enough to play in the Dutch national league and they slowly disappear from Dutch professional football. Now the difference between players born in the first quarter and players born in the last quarter should be less. The players who remain in the League all have shown enough qualities and there is no reason for a difference in their market value anymore.
17
References
Barnsley, R. H. & Thompson, A. H. (1988) Birthdate and success in minor hockey: The key to the NHL. Canadian Journal of Behavioural Science, 20, 167–176
Barnsley, R. H., Thompson, A. H. & Barnsley, P. E. (1985) Hockey success and birthdate: The RAE.
Canadian Association for Health, Physical Education and Recreation, 51, 23–28.
Barnsley, R. H., Thompson, A. H. & Legault, P. (1992) Family planning: football style. The RAE in football. International Review for the Sociology of Sport, 27, 77–88.
Bäumler, G. (1996) Der Relativalterseffekt bei Fußballspielern und seine Wechselwirkung mit dem Lebensalter. Talk at the Soccer Sport Science Symposium in Oberhaching, September 1996. Via: Musch, J. & Grondin, S. (2001) Unequal Competition as an Impediment to Personal Development: A Review of the Relative Age Effect in Sport
Doornbos, K. (1971). Date of birth and scholastic performance. Groningen: Wolters-Noordhoff. Dudink, A. (1994). Birth date and sporting success. Nature, 368, 592.
Feltz, D., & Petlichkoff, L. (1983). Perceived competence among interscholastic sport participants and dropouts. Canadian Journal of Applied Sports Sciences, 8, 231 – 235.
Grondin, S. Deshaies, P. & Nault, L.P. (1984) Quarter of birth and participation in hockey and volleyball. Journal of Sport Behavior, 27(1), 31-38.
Grondin, S., Proulx, J., & Zhou, R.-M. (1993). Date de naissance et rendement scolaire. Apprentissage
et Socialisation, 16, 169–174.
Grondin. S., & Trudeau, F. (1991). Date of birth and national hockey league: Analyses according to different parameters [in French]. Revue des Science et Techniques des Activites Physiques et
Sportives, 26, 37–45.
Malina, R. (1994). Physical growth and biological maturation of young athletes. Exercise and Sport
Sciences Reviews, 22, 389–434.
May, D., & Welch, E. (1986). Screening for school readiness: the influence of birthdate and sex.
Psychology in the Schools, 23, 100–105.
Musch, J. & Grondin, S. (2001) Unequal Competition as an Impediment to Personal Development: A Review of the Relative Age Effect in Sport. Developmental Review, 21, 147–167.
Musch, J., & Hay, R. (1999). The relative age effect in soccer: Cross-cultural evidence for a systematic discrimination against children born late in the competition year. Sociology of Sport, 16, 54– 64.
Nakata, H. & Sakamoto, K. (2011) Relative age effect in Japanese male athletes. Perceptual and
18 Roberts, G. C., Kleiber, D. A., & Duda, J. L. (1981). An analysis of motivation in children’s sport: The
role of perceived competence in participation. Journal of Sport Psychology, 3, 206–216. Vallerand, R. J., Deci, E. L., & Ryan, R. (1988). Intrinsic motivation in sport. Exercise and sport sciences
reviews, 16, 389–425.
Verhulst, J. (1992). Seasonal birth distribution of West European soccer players: A possible explanation. Medical Hypotheses, 38, 346–348.
Wendt, H. (1974). Early circannual rhythms and adult human behaviour: components of a chronobehavioural theory, and critique of persistent artifacts. International Journal of
Chronobiology, 2, 57–86.
Wendt, H.W. (1978) Season of Birth, Introversion, and Astrology: A Chronobiological Alternative, The
19
Appendix A:
Chi-Square Test
Month of birth
Observed N Expected N Residual
1 55 38,9 16,1 2 45 35,4 9,6 3 50 38,9 11,1 4 40 37,6 2,4 5 38 38,9 -,9 6 44 37,6 6,4 7 37 38,9 -1,9 8 35 38,9 -3,9 9 36 37,6 -1,6 10 28 38,9 -10,9 11 27 37,6 -10,6 12 23 38,9 -15,9 Total 458 Test Statistics Month of birth Chi-Square 26,785 Degrees of freedom 11 Asymptotic Significance ,005
20
Appendix B
Players who played for multiple teams Old club New club
Bartholomew Ogbeche Cambuur Leeuwarden Willem II
Ben Rienstra PEC Zwolle AZ
Botteghin FC Groningen Feyenoord
Etienne Reijnen Cambuur Leeuwarden FC Groningen
Guus Hupperts AZ Willem II
Jordy Buijs SC Heerenveen Roda JC
Kristoffer Peterson FC Utrecht Roda JC
Mark Diemers FC Utrecht De Graafschap
Rai Vloet PSV Cambuur Leeuwarden
Renato Tapia FC Twente Feyenoord
Rochdi Achenteh PEC Zwolle Willem II
Ruben Ligeon Willem II FC Utrecht
Stijn Wuytens Willem II AZ
Tarik Kada SC Heerenveen Heracles Almelo
21
Appendix C. Nationality of the players
Zweden 12 Zuid-Afrika 2 Zimbabwe 1 Verenigde Staten 2 Venezuela 1 Uruguay 2 Turkije 3 Tsjechië 3
Trinidad & Tobago 1
Suriname 2 Spanje 2 Slowakije 2 Slovenië 2 Sierra Leone 1 Servië 4 Schotland 1 Roemenië 1 Portugal 1 Polen 7 Peru 1 Paraguay 1 Oostenrijk 1 Oekraïne 1 Noorwegen 4 Nigeria 4 Nieuw-Zeeland 1 Nederland 284 Mexico 3 Marokko 8 Litouwen 1 Kroatië 2 Kazachstan 1 Kameroen 1 Kaapverdië 2 Japan 2 Ivoorkust 1 Israël 1 Iran 1 Ierland 1 IJsland 2 Hongarije 1 Griekenland 3 Ghana 5 Georgië 2 Frankrijk 5 Finland 1 Engeland 4 Ecuador 1 Duitsland 9 Denemarken 11 Curaçao 3 Congo DR 2 Colombia 1 China 1 Chili 1 Brazilië 5 Bosnië-Herzegovin 2 België 21 Australië 5 Aruba 1 Albanië 1 Country Freq.
